This repo include implementation of spiking differential neural network (SDNN) and examples of learning SDNN on 3 experimental datasets (winter, flights, optokinetics). All datasets contain records of the angles of rotation of the eye and head which describe the vestibulo-ocular reflex (VOR). Fig below show identification of VOR by SDNN with izhikevich activation function.
Download repo and setting up a virtual environment:
git clone https://github.com/cut4cut/spikennet.git
cd spikennet
. venv/bin/acivate
pip install -r requirements.txt
Fast start testing SDNN via console app. This command learn SDDN with sigmoidal activation function on dataset of flight's experimental data:
python main.py -a sigmoidal -d flights
After runing this commnd you can see report of model's learning. Also you can use -h to show all comands.
python main.py -h